An engine that never builds the list.
Gradient Path is a column generation engine. It searches the whole problem rather than nudging a starting point, and it is tuned to return an answer while you can still act on it.
The best plan is hiding in a very large haystack.
Every delivery round, crew roster and vehicle assignment is a choice among an enormous number of alternatives. A depot with fifty stops has more possible route combinations than there are atoms in the observable universe. No amount of computing power gets through a list that long.
So most planning software stops looking almost immediately. It applies rules of thumb, produces something workable, and moves on — leaving a gap between the plan you run and the plan you could have run. That gap is real money: vehicles that needn't have left the depot, shifts that needn't have been paid.
Planning also happens on a clock. A brilliant answer that arrives after the vehicles have left is worth nothing, which is why the practical question is never "what is the perfect plan?" but "how good a plan can we have by the time we must decide?"
That is the question we build for. Not by searching harder, but by searching differently.
Work out which plan is worth considering next.
Instead of considering every possible plan, Gradient Path works out which one is worth considering next.
-
Start small
Begin with a handful of workable routes or shift patterns — nowhere near all of them, just enough to get going.
-
Ask a precise question
Given the current plan, is there a route or roster line that would make it cheaper? The engine searches for that one candidate directly, rather than sifting a catalogue.
-
Repeat while it pays
Each new candidate improves the plan. The engine keeps going while the improvements are worth the time, and stops when they stop mattering.
Why that structure matters is worth being clear about. Most planning software improves a plan by nudging it — swap two stops, move a shift, keep whatever came out better. Methods like that can only ever find what sits close to where they happened to start, and they have no way of telling a genuinely good plan from a dead end that merely resists further tinkering.
Column generation is a different kind of method. It reasons about the problem as a whole, which means the best plan of all is always within its reach — it can arrive at answers that no amount of local tweaking would ever stumble across.
Followed to the end, that search would eventually exhaust the problem. Real operations don't have that kind of time, so we layer well-tuned heuristics on top of it — deciding which candidates are worth pursuing, which branches to abandon early, and when further searching has stopped paying for itself.
That combination is the whole idea: a method that searches globally, run in a way that is fast enough to be useful. We are not chasing the last fraction of a percent at any cost — we are after a plan good enough to change your numbers, delivered while it can still be acted on.
A global search
Reasons about the whole problem, so the best plan is always within reach — not just whatever sits near the starting point.
Answers on the clock
Built to give you the best plan it can within the time you have, not the perfect one eventually.
Industrial scale
Built for the problem sizes that arise in real operations, not textbook examples.
One engine, many problems
The same core solves vehicles, crews, aircraft and shifts. Every improvement benefits all of them.
See where it is put to work.
The same engine drives routing, scheduling, workforce and assignment problems across several industries.